Mobile Computers as Scientific Computing Machines

As pervasive mobile computing becomes more of a reality the question arises whether a mobile device such as a tablet can be sufficient to serve as the sole personal computing device of a scientist. To this end the performance of an Android based tablet computer is compared to that of a Notebook. Due to its prevalence within the scientific community and its availability on both platforms, Python is used as the basis of the comparison.The conclusion is that the basic tablet hardware platform compares well to that of the Notebook. The software also compares well in terms of usability and features but lacks performance where floating-point computations are concerned.The notebook-to-tablet ratio of performance varied from 3x for Int32 variables to 100x for float32 and float64. The slow floating-point computations were found not to be due to hardware limitations but most likely to be the result of incorrect compiler switches used when compiling the Python application for the tablet.